Eliciting User Preferences Based on Online Game

Abstract

Eliciting user preferences for large datasets and creating rankings based on these preferences has many practical applications in community based sites. This project gives a new method to elicit user preferences with an online game. In this game we do not ask users what they prefer but what a random person would prefer and rewards them if their prediction is correct. This is done through the implementation of two-player game in which initially a random player is chosen with whom the game will be played and if the random player is not available at that time, the game will be played with the computer (bot) which will appear as if the user is playing game with another human player. Then each player is shown two images and is asked to click on the image their partner would prefer. Here, the player has to think that whether he should pick the one he like best, or the one he think his partner likes best, or the one he think his partner thinks he like best, etc. The game has proven to be enjoyable, has attracted tens of thousands of people and has already collected millions of judgments. Through the judgments collected for the particular user on the basis of game played by the user we can predict the gender of the user. Thus we will show that how merely observing user preferences on a specially chosen set of images can predict a user‟s gender with high probability. This shows that responding to a request for information, such as which of two images a user prefers, can actually reveal significant information about a person. Under these circumstances, it becomes questionable whether people really can protect their privacy online. This game can collect many types of useful data from the players. It is possible to extract a global beauty ranking within a large collection of images. Also after a person has played the game on a small number of pairs of images , it is possible to extract the person‟s general